Agent safety checks: prompt-injection scan, honeypot check, x402 preflight + crypto data (USDC/x402)
Server Details
Checks before an agent acts: prompt-injection scan, honeypot check, x402 preflight. 3 free calls.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 7 tools
Each tool has a distinct primary job: input_guard/preview scan text, stealth_read fetches and cleans a page, preflight vets a merchant URL, token_risk inspects an ERC-20, shield_check aggregates a spend decision, and buy_pack_with_card handles billing. The main soft spots are input_guard vs input_guard_preview (same engine, different depth) and preflight vs shield_check, since shield_check internally calls preflight — an agent may be unsure whether to call the aggregate or the component.
All names are snake_case and mostly noun_verb or noun-phrase style (input_guard, shield_check, token_risk, stealth_read), with input_guard_preview following the same stem cleanly. buy_pack_with_card switches to a full verb_noun phrase and preflight is a bare noun, which are minor but visible deviations from the rest.
Seven tools is well within the ideal 3-15 range, and each one covers a distinct safety or billing concern rather than padding the surface. No tool looks redundant enough to be dropped except arguably the preview variant, which earns its place as a free tier.
Core checks (injection scan, page read, merchant preflight, token risk, spend firewall, credit purchase) are covered, but lifecycle gaps remain: there is no tool to create/read/update the stored spend policy that shield_check depends on, and reservation confirm/release are exposed only as raw HTTP endpoints (POST /shield/confirm, /shield/release), forcing the agent off the tool surface mid-payment. Key/balance management is also absent.
Available Tools
7 toolsbuy_pack_with_cardAngel: 5,000 calls by cardAInspect
For businesses without USDC: 5,000 calls €9.00, one-time, 90 days, businesses only, paid in EUR by card or Link. Returns checkout_url for a person to open; the checkout asks for the company or freelance name. After payment the success page (Accept: application/json) returns the key, and it is e-mailed. After: any tool with bundle_token.
Inputs: calls, email.
input_guardRemove hidden content from untrusted text and flag text aimed at the modelARead-onlyInspect
Removes hidden content and flags text aimed at the model, for untrusted input before an agent acts on it: web pages, emails, RAG chunks, tool and MCP responses. Strips invisible Unicode (tag smuggling decoded, zero-width, bidi) and HTML hidden from people; flags role takeover, forged system or tool turns, requests for keys, the prompt or payments, exfiltration links, homoglyphs and base64 payloads, in 17 languages. Rule-based. Verdict clean, caution or hostile, score, findings, cleaned text. Free with a key (500 calls) or 5 calls a day without one; then 0.02 USDC per call via x402, or 1 call from a key. After: preflight before your agent pays anything the content pointed to (the result's next field has the exact call).
Inputs: url, text, payment_tx, bundle_token, +1 more.
input_guard_previewPrompt-injection quick scan (free)ARead-onlyInspect
Free prompt-injection quick scan for untrusted input: up to 4,000 characters of text, or one https page, checked for hidden instructions aimed at agents (role override, invisible Unicode, hidden HTML) in 17 languages incl. Chinese, Japanese, Korean, Hindi, Arabic. Returns verdict, score and the top 3 findings. After: input_guard for every finding and the cleaned text.
Inputs: url, text.
preflightPre-purchase check for any shop, checkout link, payment page, API or x402 endpoint — or for a need: find the offer and check itARead-onlyInspect
Before an agent buys, pays or hires: send url (shop, checkout, payment link, API or x402 endpoint), need (what to buy in plain words; Angel takes the pick from the open sea at mareliberum.dev and checks it) or agent (mare:… resident: its trust score). One unpaid request each: TLS, domain age, look-alikes, redirects, payment provider, card fields, text aimed at agents and, for x402, USDC contract, payTo and price traps. Verdict pay, caution or do_not_pay, risk_score, reasons. Free with a key (500 calls) or 5 calls a day without one; then 0.02 USDC per call via x402, or 1 call from a key. After: input-guard on the page text you will act on (the result's next field has the exact call).
Inputs: url, need, agent, budget, +8 more.
shield_checkSpend firewall: allow, hold or block an agent payment before it happensARead-onlyInspect
Spend firewall for AI agents with a wallet or card, called before every payment: checks your rules (per-payment cap, daily budget reserved atomically, allow/block lists, chains, tokens, approval threshold, new merchants, kill switch), the merchant or checkout URL (preflight), the text that triggered the payment (input-guard: hidden instructions) and non-USDC tokens (token-risk). Returns allow, hold or block with reasons; with a stored policy the amount is reserved until confirm or release. Free with a key (500 calls) or 5 calls a day without one; then 0.02 USDC per call via x402, or 1 call from a key. After: pay, then POST /shield/confirm with the reservation_id (or /shield/release if you did not pay) (the result's next field has the exact call).
Inputs: payment (required), policy, trigger, policy_id, +5 more.
stealth_readRead any public page as clean Markdown, with hidden text removedARead-onlyInspect
Returns a public https page as clean Markdown for an agent (headings, lists, tables kept; navigation, cookie banners, scripts and everything hidden from people removed), strips invisible Unicode and flags text aimed at the model — with verdict, score, findings, title and language. JSON and text pass through. One GET as a declared bot, no login or bot-protection bypass; a blocked page is never charged. Free with a key (500 calls) or 5 calls a day without one; then 0.02 USDC per call via x402, or 1 call from a key. After: preflight before your agent buys or pays anything the page offers (the result's next field has the exact call).
Inputs: url (required), payment_tx, bundle_token, payer_address.
token_riskHoneypot and rug-pull check for an ERC-20 before swap or approveARead-onlyInspect
Honeypot and rug-pull check for an ERC-20 on Base or Ethereum before a swap, approval or signature, read straight from the contract: upgradeable proxy (EIP-1967, beacon, clone), owner powers in the bytecode (mint, pause, blacklist, fee or max-tx setters, trading switch, burn from any address), owner renounced or active and the owner's share of supply; known issuers labeled. Risk low, medium or high with reasons. Signals, not investment advice. Free with a key (500 calls) or 5 calls a day without one; then 0.02 USDC per call via x402, or 1 call from a key. After: preflight before your agent pays or signs on the site that offers this token (the result's next field has the exact call).
Inputs: token (required), chain, payment_tx, bundle_token, +1 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Pay-per-call safety guards for AI agents: injection, tool-call, signing, secret, x402-trust.
Agent security pre-checks: prompt-injection scan, MCP tool-poisoning scan, x402 endpoint check.
Check a tx, EIP-712 payload, text or Base token before your agent signs. Paid per call over x402.
Pay-per-call safety checks for AI agents: screen a crypto address or URL before you transact.
Related MCP Servers
- AlicenseAqualityCmaintenanceSafety checks an agent runs before it acts — flags malicious packages, destructive shell commands, secret leaks, and web-backend holes before execution.828 PyPIMIT
- AlicenseAqualityBmaintenanceEnables AI agents to run a free fail-closed preflight gate on x402 payment challenges and transactions before signing, plus five USDC-paid on-chain truth checks covering arbitrage profitability, token safety, swap routing, quote validation, and trade sizing. Each check returns a verdict with pinned-block evidence and a verifiable receipt.49Apache 2.0
- AlicenseAqualityDmaintenancesecurity tools for AI agents: URL safety scanning, prompt injection detection (200+ patterns), email/password breach checks via HIBP, domain & IP reputation analysis, and AI skill supply chain scanning. Free tier (3 calls/day) or pay-per-request with USDC micropayments via x402.919 npm1MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to run pre-purchase safety checks on shops, checkout and payment links, paid APIs, and x402 endpoints before paying, returning a pay, caution, or do_not_pay verdict with reasons.MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.